Riverstone Networks RS 2100 Spezifikationen PDF herunterladen (Seite 17)

Fibre Channel Over IP

17Continuous Access and Data Replication Manager SAN Extensions Reference Guide

Click Technical documentation to be directed to the application notes, or

click Related information to find the tool.

Because of similarities in the method of replication, this tool is also useful in

establishing an initial estimate of the performance needs of Continuous

Access XP.

■ For DRM environments, use the DRM Performance Estimator. Its use is

documented in the Compaq SANworks Data Replication Manager Inter-site

Link Performance Analyzer White Paper. The tool and white paper are located

in the manuals page at

http://h18000.www1.hp.com/products/sanworks/drm/documentation.html

Bandwidth considerations

When sites are located many miles apart, the speed of light through fiber can

cause unacceptable delays in the completion of an I/O transaction. Increasing the

amount of available bandwidth cannot solve this problem. Careful consideration

must be given to these factors when matching your needs and wants to a particular

application. For a discussion of distance and performance in design

considerations, especially in relation to data replication, refer to the HP

StorageWorks Continuous Access EVA Performance Estimator User Guide, or the

HP StorageWorks Data Replication Manager HSG80 ACS Version 8.7P

Configuration Guide. The discussion for ACS 8.7 also applies for ACS 8.8.

If an application already exists, there is a way to more precisely determine the

needed bandwidth. This is to measure the amount of new or changed data that is

created by that application. For example:

1. Collect the peak read and write workloads for a given period of time. If using

a Windows based operating system use a tool like PERFMON to capture the

current performance requirements without CA. Likewise a similar tool exists

for other operating systems. At each sample interval, capture reads per

second (IOPS), read throughput per second (MBytes per second), writes per

second (IOPS) and write throughput per second (MBytes per second. If

possible also collect read and write latency. Do this by application, capturing

the data for each logical unit (device) used by that application.

2. Once the data is collected, create a graph of each data set so that you can see

where the peaks occur during the day. This is critical to understanding if the

peaks are at the same time or not. The graph also helps to show if the daily

average change rate is really level or very bursty. Also consider growth in

these numbers for the next 12 to 18 months. The results of this scaling

process now becomes your design goal.

1 2 ... 12 13 14 15 16 17 18 19 20 21 22 ... 119 120

Keine Kommentare

Riverstone Networks RS 2100 Spezifikationen Seite 17